DYB-PlanktonNet is a dataset contains marine plankton and suspension particles ROI images recorded from the Daya Bay (DYB), an inner bay of the South China Sea close to Shenzhen City, China.
All the images in this dataset were captured by an underwater dark-field imager system developed by Dr. Jianping Li’s research group at Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences. The images were labeled through a “human-machine mutually-assisted” process and an “amateur-sorting first, expert-verification follows” strategy to achieve both high annotation efficiency and quality. At present, the dataset has totally 46,804 ROI images in 90 categories, in which 82 are plankton, 7 are suspending particles and 1 is bubbles. The dataset is expected to contribute to the development of automated and intelligent analysis methods for marine plankton and suspending particle images classification and evaluation, hence deepen our understanding of the marine ecosystem.
- How to use?
Download and unzip the DYB-PlanktonNet.zip file, you will get 90 subfolders under the DYB-PlanktonNet folder. Each subfolder corresponds to one plankton or particle category and contains ROI images of that category.
- Images info.
All the images are color ROIs captured from Dapeng Cove between the years from 2018 to 2021. The images are saved in .jpg format with R, G, B channels, and each pixel corresponds to an actual length of 7.38μm. Standard CLAHE image processing is applied on each image to enhance its contrast. Most of the ROIs were saved with their names indicating the recording time and date information.
For the detailed dataset information, please refer to the documentation file "DYB-PlanktonNet Documentation.pdf".